Loading ...
Sorry, an error occurred while loading the content.

Finding RDF services (or: imminent death of usenet predicted)

Expand Messages
  • Dan Brickley
    On Fri, 29 Sep 2000, Dan Brickley wrote: [everything except the point he originally intended] I wanted to mention an idea I ve sat on for a few months, but
    Message 1 of 9 , Sep 29, 2000
    • 0 Attachment
      On Fri, 29 Sep 2000, Dan Brickley wrote:

      [everything except the point he originally intended]

      I wanted to mention an idea I've sat on for a few months, but failed to
      flatten.

      RDF/XML Service Discovery seems to be a hot topic. Well, XML stuff in
      particular, industry having gone bananas over it. I have in mind a
      lightweight, distributed, scaleable, standards-based infrastructure to
      support XML service and datasource discovery for B2B and P2P apps,
      middleware and n-tiered acronym generation.

      Let's start a Usenet newsgroup.

      I started to draft a request for news:alt.rdf.* group(s). Idea was
      that a simple good old fashioned usenet group would be a great way for XML/RDF
      apps to publish and discover services and data. We already have global
      infrastructure in place to support this, Perl, Java etc client libraries
      and lots more. Unlike Web-based aggregation and 'e-market' sites, the
      lack of central management depoliticises things rather helpfully. No one
      WWW site gets to be the central whiteboard on which RDF/XML apps
      scribble messages to one another. Contrast: if I'd just suggested that
      everyone uses www.danbri.org as a central service for sharing RSS files,
      that'd be rather different.

      There would be a lot of detail to work out, and I don't want to distract
      RSS-DEV from the 1.0 effort, but thought it was about time to make this
      thought bookmarkable. I lack familiarity with the politics and ritual
      of Usenet group creation, but intend to give it a go. My understanding
      is that the alt.* hierarchy is easier to get a group established in, but
      less well replicated. I suggest we (if anyone agrees this isn't
      crazy) try an alt group first, then take things from there. One
      oddity: usenet to date seems split between human-oriented text groups
      and the 'binary' groups for images etc. What I'm proposing is a
      machine-oriented group. I'm not sure if the Usenet crowd would like
      this, so eg. we might suggest that postings are multipart, with human
      and machine readable sections.

      Anyway, what do you all reckon? am I making any sense?

      Dan

      ps. would also be a nice place to experiment with digitally signed
      metadata, another sideline I'm investigating.
    • Aaron Swartz
      ... So, do you think it would be proper for us to (suggest? recommend?) the use of LINK REL= rss or LINK RSS= rss1.0p to refer to RSS files? ... What
      Message 2 of 9 , Sep 29, 2000
      • 0 Attachment
        Dan Brickley <daniel.brickley@...> wrote:

        >> I'd like to ask if someone on the list who is more knowledgeable than I could
        >> explain more about the use of LINK REL, and specifically, how one identifies
        >> that their site provides an RSS file.
        > There's not a lot to know really. What little conventions exist are in
        > the HTML 4+ specs somewhere. Everything else is made up.

        So, do you think it would be proper for us to (suggest? recommend?) the use
        of LINK REL="rss" or LINK RSS="rss1.0p" to refer to RSS files?

        > As an aside, also worth mentioning that HTML LINK types can be used
        > (per document and per hyperlink) for sitemap-style apps, which is
        > pretty close to RSSville.
        > [[ oh, just noticed Aaron's just found the link I was about to
        > mention! ;-]]

        What synergy! I definitely think that RSS is a perfect application (along
        with XHTML/HTML) for these types of links. RSS publications of weblogs could
        have so much more semantic meaning if the links used this kind of annotation
        (tying back to your demonstration of RDF-in-Weblogs). Not to mention the
        stuff being discussed over on XML-DIST-APP about
        XML-push-email-notification-etc. Whoa, lots of things connecting here.

        > I'm hopelessly overcommitted and unlikely to find much time
        > in the next month or so to work on linktypes for sitemapping.

        Had a feeling... That's OK, perhaps there are others out there willing to
        take on the task. (I can moderate, but I doubt I'm able or qualified to do
        the actual work myself.)

        --
        Aaron Swartz |"This information is top security.
        <http://swartzfam.com/aaron/>| When you have read it, destroy yourself."
        <http://www.theinfo.org/> | - Marshall McLuhan
      • Aaron Swartz
        ... I think this would be rather cool. So how would it work? People d post XML files to the newsgroup? Count me on the bandwagon! -- Aaron Swartz
        Message 3 of 9 , Sep 29, 2000
        • 0 Attachment
          Dan Brickley <daniel.brickley@...> wrote:

          > Anyway, what do you all reckon? am I making any sense?

          I think this would be rather cool. So how would it work? People'd post XML
          files to the newsgroup? Count me on the bandwagon!

          --
          Aaron Swartz |"This information is top security.
          <http://swartzfam.com/aaron/>| When you have read it, destroy yourself."
          <http://www.theinfo.org/> | - Marshall McLuhan
        • Dan Brickley
          ... Not sure enough of the rules/conventions, but seems a good approach to me. Would be nicer if it allowed a URI. Maybe the notion of an an HTML profile, see
          Message 4 of 9 , Sep 29, 2000
          • 0 Attachment
            On Fri, 29 Sep 2000, Aaron Swartz wrote:

            > Dan Brickley <daniel.brickley@...> wrote:
            >
            > >> I'd like to ask if someone on the list who is more knowledgeable than I could
            > >> explain more about the use of LINK REL, and specifically, how one identifies
            > >> that their site provides an RSS file.
            > > There's not a lot to know really. What little conventions exist are in
            > > the HTML 4+ specs somewhere. Everything else is made up.
            >
            > So, do you think it would be proper for us to (suggest? recommend?) the use
            > of LINK REL="rss" or LINK RSS="rss1.0p" to refer to RSS files?

            Not sure enough of the rules/conventions, but seems a good approach to
            me. Would be nicer if it allowed a URI. Maybe the notion of an
            an HTML profile, see view/src of http://www.w3.org/ for pointer to
            http://www.w3.org/2000/08/w3c-synd/# for an example. Would need reading
            up on.


            >
            > > As an aside, also worth mentioning that HTML LINK types can be used
            > > (per document and per hyperlink) for sitemap-style apps, which is
            > > pretty close to RSSville.
            > > [[ oh, just noticed Aaron's just found the link I was about to
            > > mention! ;-]]
            >
            > What synergy! I definitely think that RSS is a perfect application (along
            > with XHTML/HTML) for these types of links. RSS publications of weblogs could
            > have so much more semantic meaning if the links used this kind of annotation
            > (tying back to your demonstration of RDF-in-Weblogs). Not to mention the

            Don't think I circulated that yet. I'll clean it up first... (I'm not
            used to weblogging -- it feels so naked ;-)

            > stuff being discussed over on XML-DIST-APP about
            > XML-push-email-notification-etc. Whoa, lots of things connecting here.

            Everything's connected :-)

            >
            > > I'm hopelessly overcommitted and unlikely to find much time
            > > in the next month or so to work on linktypes for sitemapping.
            >
            > Had a feeling... That's OK, perhaps there are others out there willing to
            > take on the task. (I can moderate, but I doubt I'm able or qualified to do
            > the actual work myself.)

            that'd be good. It's mostly a dull matter of reading specs, and the
            interesting work of doing a literature review to see what's been done
            before. There's a strong HCI aspect to doing this right...

            Dan
          • Dan Brickley
            ... People and/or machines. Writing and reading. BTW I realise there are some sensitivities surrounding mechanical posting to Usenet, but that s the basic
            Message 5 of 9 , Sep 30, 2000
            • 0 Attachment
              On Fri, 29 Sep 2000, Aaron Swartz wrote:

              > Dan Brickley <daniel.brickley@...> wrote:
              >
              > > Anyway, what do you all reckon? am I making any sense?
              >
              > I think this would be rather cool. So how would it work? People'd post XML
              > files to the newsgroup? Count me on the bandwagon!

              People and/or machines. Writing and reading. BTW I realise there are
              some sensitivities surrounding mechanical posting to Usenet, but that's
              the basic idea.

              It generalises beyond RSS (and RDF I guess) -- you could use it for
              discovery of SOAP/XML-RPC interfaces and the like. My guess is that such
              a system would only really become interesting when the messages were
              digitally signed... Eg. I could filter for just those RDF announcements
              that were assured in some way by a listmember of rss-dev, if I had a database
              of the public keys of folk on this and similar lists(*).

              Anyone who knows news:comp.infosystems.announce should get the
              picture. An alt.* group rather than a comp.infosystems.* group seems the
              best place to begin experimenting though.

              Dan


              (*)
              BTW following this tangent... a few weeks back Rael, Edd,
              myself and others made a start on a mutual key-signing effort, mostly
              using http://www.gnupg.org/ -- that might provide some basis for
              dealing with the chaos that a news:alt.rdf.* infospew might create. I
              started work on an RDF representation of those key-signings, but fear
              what I've done to date is rather toy. (This is a small part of an
              insanely ambitious project I'm working on, but won't shout about here
              until my code matches my rhetoric ;-)

              For the curious,
              http://snowball.ilrt.bris.ac.uk/~pldab/rdfweb/danbri.wot.rdf
              graphically depicted in:
              http://snowball.ilrt.bris.ac.uk/~pldab/rdfweb/danbri.wot.gif
              shows a fragment of this data, dumped from GPG into RDF and graphed with
              my RDFViz service.
            • Libby Miller
              Hi, I work with Dan Brickley at the ILRT and I ve been following the RSS1.0 discussions with interest. Ages ago I looked at some sitemaps
              Message 6 of 9 , Oct 2, 2000
              • 0 Attachment
                Hi,

                <de-lurk>
                I work with Dan Brickley at the ILRT and I've been following the RSS1.0
                discussions with interest.
                </de-lurk>

                Ages ago I looked at some sitemaps stuff with Dan - I'd be quite keen to
                do a literature review. I'll see how things pan out later this week.

                Libby


                On Fri, 29 Sep 2000, Dan Brickley wrote:

                > On Fri, 29 Sep 2000, Aaron Swartz wrote:
                >
                > > Dan Brickley <daniel.brickley@...> wrote:
                > >
                > > >> I'd like to ask if someone on the list who is more knowledgeable than I could
                > > >> explain more about the use of LINK REL, and specifically, how one identifies
                > > >> that their site provides an RSS file.
                > > > There's not a lot to know really. What little conventions exist are in
                > > > the HTML 4+ specs somewhere. Everything else is made up.
                > >
                > > So, do you think it would be proper for us to (suggest? recommend?) the use
                > > of LINK REL="rss" or LINK RSS="rss1.0p" to refer to RSS files?
                >
                > Not sure enough of the rules/conventions, but seems a good approach to
                > me. Would be nicer if it allowed a URI. Maybe the notion of an
                > an HTML profile, see view/src of http://www.w3.org/ for pointer to
                > http://www.w3.org/2000/08/w3c-synd/# for an example. Would need reading
                > up on.
                >
                >
                > >
                > > > As an aside, also worth mentioning that HTML LINK types can be used
                > > > (per document and per hyperlink) for sitemap-style apps, which is
                > > > pretty close to RSSville.
                > > > [[ oh, just noticed Aaron's just found the link I was about to
                > > > mention! ;-]]
                > >
                > > What synergy! I definitely think that RSS is a perfect application (along
                > > with XHTML/HTML) for these types of links. RSS publications of weblogs could
                > > have so much more semantic meaning if the links used this kind of annotation
                > > (tying back to your demonstration of RDF-in-Weblogs). Not to mention the
                >
                > Don't think I circulated that yet. I'll clean it up first... (I'm not
                > used to weblogging -- it feels so naked ;-)
                >
                > > stuff being discussed over on XML-DIST-APP about
                > > XML-push-email-notification-etc. Whoa, lots of things connecting here.
                >
                > Everything's connected :-)
                >
                > >
                > > > I'm hopelessly overcommitted and unlikely to find much time
                > > > in the next month or so to work on linktypes for sitemapping.
                > >
                > > Had a feeling... That's OK, perhaps there are others out there willing to
                > > take on the task. (I can moderate, but I doubt I'm able or qualified to do
                > > the actual work myself.)
                >
                > that'd be good. It's mostly a dull matter of reading specs, and the
                > interesting work of doing a literature review to see what's been done
                > before. There's a strong HCI aspect to doing this right...
                >
                > Dan
                >
                >
                > To unsubscribe from this group, send an email to:
                > rss-dev-unsubscribe@egroups.com
                >
                >
                >
                >
              • h4x0rific
                I stumbled across this thread via Google today - did anything ever come of the Usenet as RDF exchange idea ? Was the proposal rejected ? I apologize for
                Message 7 of 9 , Dec 20, 2001
                • 0 Attachment
                  I stumbled across this thread via Google today - did anything ever
                  come of the Usenet as RDF exchange idea ? Was the proposal rejected ?

                  I apologize for dredging the archives, but the idea seems very
                  interesting.

                  :D

                  -brian donovan
                  http://www.lophty.com/kumo/

                  --- In rss-dev@y..., Dan Brickley <daniel.brickley@b...> wrote:
                  > Let's start a Usenet newsgroup.
                  >
                  > I started to draft a request for news:alt.rdf.* group(s). Idea was
                  > that a simple good old fashioned usenet group would be a great way
                  for XML/RDF
                  > apps to publish and discover services and data. We already have global
                  > infrastructure in place to support this, Perl, Java etc client libraries
                  > and lots more. Unlike Web-based aggregation and 'e-market' sites, the
                  > lack of central management depoliticises things rather helpfully. No one
                  > WWW site gets to be the central whiteboard on which RDF/XML apps
                  > scribble messages to one another. Contrast: if I'd just suggested that
                  > everyone uses www.danbri.org as a central service for sharing RSS files,
                  > that'd be rather different.
                Your message has been successfully submitted and would be delivered to recipients shortly.